BackgroundThe relationship between shot-to-shot sampling rates for dynamic myocardial computed tomography perfusion (CTP) and robustness of CTP-derived myocardial blood flow (CT-MBF) is debatable. We retrospectively investigated the influence of a reduced sampling rate for dynamic CTP on CT-MBF computation and diagnostic performance for detecting myocardial perfusion abnormalities. MethodsPharmacological stress dynamic whole-heart CTP was performed in 120 patients suspected with coronary artery disease. Dynamic CTP data were obtained for 30 continuous heartbeats during the R-peak to R-peak (1RR) interval on electrocardiography. Three additional datasets were created with sub-sampling acquisitions every 2, 3, and 4 heartbeats from the original dataset as interval times of 2RR, 3RR, and 4RR, respectively. CT-MBF was calculated using deconvolution analysis and determined as the mean value of the whole heart (global CT-MBF) and using the 16-segment model (segmental CT-MBF). The diagnostic performance of segmental CT-MBF for detecting perfusion abnormalities was compared to that of cardiac magnetic resonance imaging as the gold standard in 32 of 120 patients. These results were compared among the four CTP datasets. ResultsGlobal CT-MBFs for 1RR, 2RR, 3RR, and 4RR sampling were 1.57 ± 0.34, 1.54 ± 0.36, 1.51 ± 0.37, and 1.41 ± 0.33 mL/g/min, respectively. Areas under the receiver operating characteristic curves of segmental CT-MBF for 1RR, 2RR, 3RR, and 4RR sampling were 0.84, 0.83, 0.79, and 0.76, respectively (1RR versus [vs.] 2RR, non-significant; 1RR vs. 3RR or 4RR, p < 0.05). ConclusionCT-MBF with 2RR sampling has similar performance with regard to quantification and detecting myocardial perfusion abnormalities as that with 1RR sampling.